Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Gautam Kamath

Gautam Kamath

4 indexed papers

Recent (6 mo)
4
With code
0
Influential cites
0
Benchmarked
0

Publications per year

4
26

Top categories

ML×4Crypto×4NLP×2

Frequent co-authors

Peihan Liu2×
Lucas Rosenblatt2×
Weiwei Kong2×
Natalia Ponomareva2×
Rachel Cummings2×
Roxana Geambasu2×

Research Timeline

2026
ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities?

The paper introduces ContinuousBench, a novel benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge, finding that state-of-the-art DP synthesis methods generally fail to achieve this capability gain.

ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities?

The paper introduces ContinuousBench, a dynamic benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge and capabilities from sensitive source corpora, finding that current state-of-the-art DP methods generally fail to achieve this.

Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses

The paper establishes tight upper and lower bounds on the statistical cost of approximate machine unlearning for smooth strongly convex losses, showing that the optimal unlearning rate depends critically on the relationship between the unlearning parameter $\varepsilon$ and the model dimension $d$.

Sequential Data Poisoning in LLM Post-Training

The paper introduces the threat model of sequential data poisoning, demonstrating that multiple, collaborating attackers can exploit compound vulnerabilities in LLM post-training pipelines that are invisible when analyzing individual stages.

Highlighted terms show continued research focus across papers

Papers

cs.LGcs.CRRecentJun 3, 2026

Sequential Data Poisoning in LLM Post-Training

Jack Sanderson, Yihan Wang, Xiaoqian Lu, Gautam Kamath +1 more

The paper introduces the threat model of sequential data poisoning, demonstrating that multiple, collaborating attackers can exploit compound vulnerabilities in LLM post-training pipelines that are in…

View →
cs.LGcs.CLcs.CRRecentJun 1, 2026

ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities?

Peihan Liu, Lucas Rosenblatt, Weiwei Kong, Natalia Ponomareva +6 more

The paper introduces ContinuousBench, a novel benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge, finding that state-of-the-art DP…

View →
cs.LGcs.CLcs.CRRecentJun 1, 2026

ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities?

Peihan Liu, Lucas Rosenblatt, Weiwei Kong, Natalia Ponomareva +6 more

The paper introduces ContinuousBench, a dynamic benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge and capabilities from sensitive…

View →
cs.LGcs.CRRecentJun 1, 2026

Near-Optimal Pure Machine Unlearning for Smooth Strongly Convex Losses

Matthew Regehr, Gautam Kamath, Andrew Lowy

The paper establishes tight upper and lower bounds on the statistical cost of approximate machine unlearning for smooth strongly convex losses, showing that the optimal unlearning rate depends critica…

View →